Phytochemistry 72 (2011) 14661472

Contents lists available at ScienceDirect

Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem

Characterisation of the arabinose-rich carbohydrate composition of immature

and mature marama beans (Tylosema esculentum)
Minah M. Mosele a,b, se S. Hansen a, Sren B. Engelsen a, Jerome Diaz c, Iben Srensen c, Peter Ulvskov c,
William G.T. Willats c, Andreas Blennow c, Jesper Harholt c,
a

Quality and Technology, Department of Food Science, Faculty of Life Sciences, University of Copenhagen, Rolighedsvej 30, 1958 Frederiksberg C, Denmark
Department of Extension and Training, National Food Technology Research Center, Mpuutsane Industrial Area, Private Bag 008, Kanye, Botswana
Section for Plant Glycobiology, VKR Research Centre Pro-Active Plants, Department of Plant Biology and Biotechnology, Faculty of Life Sciences, University of Copenhagen,
Thorvaldsensvej 40, 1871 Frederiksberg C, Denmark
b
c

a r t i c l e

i n f o

Article history:
Received 28 October 2010
Received in revised form 1 February 2011
Available online 3 May 2011
Keywords:
Marama bean
Morama bean
Tylosema esculentum
Legume
Starch
Soluble sugars
Cell wall composition
Pectin
CoMPP
Linkage analysis
HPAEC-PAD

a b s t r a c t
Marama bean (Tylosema esculentum) is an important component of the diet around the Kalahari Desert in
Southern Africa where this drought resistant plant can grow. The marama bean contains roughly 1/3 proteins, 1/3 lipids and 1/3 carbohydrates, but despite its potential as dietary supplement little is known
about the carbohydrate fraction. In this study the carbohydrate fraction of immature and mature
marama seeds are characterised. The study shows that the marama bean contains negligible amounts
of starch and soluble sugars, both far less than 1%. The cell wall is characterised by a high arabinose content and a high resistance to extraction as even a 6 M NaOH extraction was insufcient to extract considerable amounts of the arabinose. The arabinose fraction was characterised by arabinan-like linkages and
recognised by the arabinan antibody LM6 and LM13 indicating that it is pectic arabinan. Two pools of
pectin could be detected; a regular CDTA (1,2-diaminocyclohexane-N,N,N0 ,N0 -tetraacetic acid) or enzymatically extractable pectin fraction and a recalcitrant pectin fraction containing the majority of the arabinans, of which about 40% was unextractable using 6 M NaOH. Additionally, a high content of mannose
was observed, possibly from mannosylated storage proteins.
2011 Elsevier Ltd. All rights reserved.

1. Introduction
Marama bean (Tylosema esculentum Burchell A. Schreiber) commonly called morama bean, tsin bean or gemsbok bean is an important component of the diet in settlements around the Kalahari
Desert (National Research Council, 1979). It is a desiccant-tolerant
plant with an ability to grow in high temperatures and dry environments such as the Kalahari area. Raw mature seeds of marama beans
store well and remain edible for years (National Research Council,
1979). Leguminous seeds are an important part of the diet of rural
communities in developing countries as they provide proteins, lipids
and carbohydrates (Ketshajwang et al., 1998).
The mature seeds of T. esculentum are encapsulated in woody
pods with 12 seeds and the pods open at maturity. Their seed
coats, which are removed before consumption, are reddish to
brownish black in colour (National Research Council, 1979). Previous work of mature marama seeds (Bower et al., 1988; Amarteio

Corresponding author. Tel.: +45 35333306; fax: +45 35333300.

E-mail address: jesh@life.ku.dk (J. Harholt).
0031-9422/$ - see front matter 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.phytochem.2011.03.027

and Moholo, 1998; Holse et al., 2010; Mosele et al., 2011) has indicated that they are a rich source of proteins and lipids, both above
30%, making them comparable to soya bean and peanut. They also
have a considerable content of dietary bre (1927%), and mineral
content is similar to that of peanut and approaching that of soya
bean (Holse et al., 2010). The immature seeds are rich in proteins
containing around 21% (w/w) (Mosele et al., 2011). Other studies
have focused on the quality of marama bean oil (Mitei et al.,
2008) and characterisation of the fatty acids, phytosterols and vitamin E compounds in the oil (Mitei et al., 2009).
Despite the potential of marama bean as a healthy nutritive
crop for developing countries none of its carbohydrates have been
thoroughly studied. Potential use of the carbohydrate fractions in
food applications could yield valuable income for rural communities gathering marama beans. Studies on the use of pressed marama oil have been initialised and the defatted press rest might be
readily available for potential extraction of useful carbohydrates.
Carbohydrates, especially polysaccharides are important in the
food industry because they can be used as thickeners, stabilisers,
texturisers and gelling agents (Viarta et al., 2006; Khurana and
Kanawjia, 2007). A number of studies have investigated the

M.M. Mosele et al. / Phytochemistry 72 (2011) 14661472

composition of carbohydrates in legume seeds, with special

emphasis on starch and soluble sugars. Lentils and chickpeas contain around 50% (w/w) starch and approximately 20% (w/w) dietary bre (Aguilera et al., 2009a). In soya bean, which compares
to marama bean in nutritional quality, the carbohydrate components include 5.3% oligosaccharides, 2.45.5% polysaccharides
(crude bre) and 0.20.9% starch (Reddy et al., 1984). The main soluble sugar in soya bean is sucrose, followed by stachyose (Hou
et al., 2009; Saldivar et al., 2011). The total carbohydrate content
(by difference) of marama bean has been reported to constitute between 18.9% and 24.1% but has never been studied in detail (Bower
et al., 1988; Amarteio and Moholo, 1998; Mosele et al., 2011).
Structure elucidation of complex polysaccharides can be challenging due to the inherent heterogeneity and due to the differences in extractability. Analysis of legume seed cell wall is no
exception and high protein content in the cell wall fraction does
not make its structure elucidation any easier. Holse et al. (2011)
characterised the bulk carbohydrate content of intact mature marama bean using different spectroscopic techniques. However a detailed characterisation of the cell wall polysaccharides was not
possible without fractionation.
Dicot cell walls are generally considered to consist of three fractions, namely pectin, hemicellulose and cellulose. Pectin can be
fractionated into homogalacturonan (HG), rhamnogalacturonan I
(RG I), rhamnogalacturonan II (RG II) and xylogalacturonan
(XGA). Apparently, these polymers are covalently linked to each
other but it has proven very difcult to obtain unambiguous information on how the different pectic polysaccharides are connected
wherefore several models exist (Vincken et al., 2003). Recent reviews have tried to elucidate the relationship of these pectic polysaccharides (Caffall et al., 2009; Scheller et al., 2007). RG I is built
up of a backbone containing the disaccharide (a-1?4-GalA-a1?2-Rha) as the basic repeating unit. The rhamnosyl residues
can be substituted with galactan, arabinan or arabinogalactan side
chains. The galacturonic acid (GalA) residues can furthermore be
acetylated as in homogalacturonan. In some species the arabinose
and galactose residues in RG I side chains can be substituted with
ferulic and coumaric acid esters (Fry, 1982).
The aim of this study is to characterise the carbohydrate composition of marama bean seeds at the two developmental stages:
immature and mature, corresponding to the stages utilised for
consumption.

2. Results and discussion

2.1. Starch
Marama bean is virtually devoid of starch in both immature and
mature seeds (Table 1). On average, the starch content is 0.2% dry
mass, similar to that of mature soya bean as reported by Reddy
et al. (1984) at 0.20.9%; Karr-Lilienthal et al. (2005) at 0.5%, and
Saldivar et al. (2011) at 0.21.0%. Holse et al. (2011) also found
starch signals to be absent in the infra red and NMR spectra of mature marama seeds. Carob seeds (Ceratonia siliqua), which is in the
same family as marama bean also has a low starch content of 0.1%
(Avallone et al., 1997). Other legumes such as pea, lentil and chick-

Table 1
Starch content of marama bean cotyledon from immature and mature seeds (% dry
mass).
Sample

Starch content %

Immature seeds
Mature seeds

0.19 (0.003)
0.17 (0.003)

Values in parentheses indicate standard deviations, n = 3.

1467

pea have a starch content of more than 40% (Dalgetty and Baik,
2003). Aguilera et al. (2009a) showed that total starch is 53.4% in
chickpea (Cicer arietinum L.) and 46.3% in lentil (Lens culinaris L.).
The negligible amount of starch made detailed characterisation
of starch, e.g. amylose/amylopectin ratio, unwarranted.
2.2. Soluble sugars
In general the content of soluble sugars is very negligible (less
than 1% dry mass) in both immature and mature seeds. The main
soluble sugar in immature seeds is glucose, followed by myoinositol and fructose (Table 2). Mature seeds had a higher amount of sucrose, followed by myoinositol and rafnose. The presence of
rafnose in the mature seeds conrm the observations by Holse
et al. (2011) using 1H HR-MAS NMR.
In a study conducted by Aguilera et al. (2009b) the main soluble
carbohydrate found in white beans and pink-mottled cream beans
was also sucrose. However, other legumes contain rafnose family
oligosaccharides (RFOs) as the main soluble sugars. The main soluble carbohydrate in soya bean (Glycine max), dolichos (Lablab purpureus) and cowpea (Vigna unguiculata) is RFOs (mainly stachyose),
followed by sucrose, while jack bean (Canavalia ensiformis) has
RFOs rafnose and ciceritol, followed by sucrose (Martn-Cabrejas
et al., 2008).
A pattern of accumulation characterised by glucose, myoinositol, fructose, trehalose and ribose was observed at the immature
stage, which disappeared at the mature stage as these sugars further decreased or were no longer detectable. The content of arabinose, rafnose, sucrose and maltose increased with maturation.
Overall, a marked difference was observed between the two samples, with immature seeds having a higher amount of soluble sugars. However, soluble sugars represent a negligible percentage of
total carbohydrates, as also observed for other legumes (Reddy
et al., 1984).
2.3. Alcohol insoluble residue (AIR)
The composition of the non-cellulosic monosaccharides in AIR
of mature marama seeds was characterised by a high content of
arabinose, and with lower levels of mannose and galactose
(Fig. 1). The immature seeds had a high content of mannose, followed by intermediate amounts of glucose and galactose. The
mannose in the linkage analysis described below only show linkages characteristic for protein mannosylation indicating that the
mannose in both samples originates from protein glycosylation
(Lis and Sharon, 1978). As an abundance of cytoplasmic protein
bodies have been visualised in marama seeds, possibly containing
the majority of the protein associated mannose, the mannose detected in the sugar composition analysis of the non-cellulosic
monosaccharides in AIR should not be considered part of the cell
wall (Mosele et al., 2011). High mannose content has previously
been reported in legume seeds with no evidence for its origin (Mullin and Xu, 2000).
Generally, legume seed cell wall contains high levels of galactose and arabinose, supposedly from long side chains of RG I (Huisman et al., 1998; Dalgetty and Baik, 2003). The presence of
xyloglucan, mannan and callose (b-1,-3 glucan) was conrmed in
CoMPP analysis (Fig. 2). However, based on the composition analysis of AIR the quantities of these polymers are minor compared to
pectin. Both arabinogalactan protein (AGP) and extensin were also
detected in the CoMMP analysis and with similar extraction prole
as seen in Arabidopsis (Fig. 2) (Moller et al., 2007). Xylan could not
be detected in this analysis and xylogalacturonan as recognised by
the LM8 antibody, known to recognise an epitope present in a subpool of xylogalacturonans (Jensen et al., 2008; Nakamura et al.,
2002) was not detected using CoMPP.

1468

M.M. Mosele et al. / Phytochemistry 72 (2011) 14661472

Table 2
Soluble sugar content of marama bean cotyledon from immature and mature seeds (% dry mass).
Soluble sugar

Immature (nmol/g)

Mature (nmol/g)

Immature (ng/mg)

Mature (ng/mg)

Myoinositol
Trehalose
Arabinose
Glucose
Fructose
Ribose
Sucrose
Rafnose
Maltose

64.1 (0.7)
1.6 (0.2)
18.2 (1.8)
274.8 (41.0)
55.8 (15.0)
13.0 (1.2)
11.0 (2.4)
1.9 (0.8)
2.2 (0.6)

49.3 (2.1)
n/d
30.0 (1.0)
n/d
n/d
2.1 (1.8)
123.5 (2.1)
27.0 (3.7)
15.4 (0.6)

11.5 (0.1)
0.6 (0.1)
2.7 (0.3)
49.5 (7.4)
10.1 (2.7)
1.9 (0.2)
3.8 (0.8)
1.0 (0.4)
0.8 (0.2)

8.9 (0.4)
n/d
4.5 (0.1)
n/d
n/d
0.3 (0.3)
42.3 (0.7)
13.6 (1.9)
5.3 (0.2)

n/d, not detected.

Values in parentheses indicate standard deviations, n = 3.

250

Immature
Mature

g
g/mgg AIR
R

200

150

100

50

0
Fucose

Rhamnose Arabinose Galactose Glucose

Xylose

Mannose

GalA

Sugars
Fig. 1. Sugars (lg/mg) of the alcohol insoluble residue (AIR) of immature (unripe)
and mature (fully ripe) marama seeds hydrolysed by TFA (GalA = galacturonic acid).

b JIM
Exxtenssin (m
M 200)
mAb
Exxtennsin (mA
M 1))
( Ab LM
A
AGP
P (mA
Ab JIM
J 13)
AGP
P (m
mAb JIM 8)
AGP
P (m
mAb LM 2)
BM 3a)
C
Celluulosee (CB
(1-3))-gluucan (mA
Ab BS
B 4000-2
2)

(1-3
mAb
3)(1--4)-gglucaan (m
b BS
S 4000-3)
(11-4)--man
Ab BS
B 400-4
4 4)
nnann (mA
(1--4)-aarabiinoxy
Ab LM
L 11)
1
ylann (mA

Ab LM
L 10)
(1-44)-xxylan
n (mA
X
XyG ((+fuuc., mAb
m b CCRC--M1))
X
XyG
G (mA
Ab LM
L 15)
Xyylogalacturonan (mA
Ab LM
L 88)
Arrabinnan ((mAb
M 13
3)
b LM

(1-55)-araabinnan (mAb
b LM
M 6)

(1-44)-gaalacttan (m
b LM
M 5)
mAb
H
HG ((Higgh D
M 7)
DE, mAb
m JIM
H
HG (Low
M 5)
w DE
E, mAb
m JIM
CDTA

89 36 29 28 35 0

7 53 35 21 9 12

4M NaOH 24 0 72 45 45 0 76 100 0 0 35 8 15 49 11 10 9 0 5

Cadoxen

12 0 37 18 18 0 25 33 0

0 20 10 12 36 6

Fig. 2. Relative abundance of AIR cell wall components from mature marama seeds
as judged by CoMMP analysis displayed as a heat map (HG = homogalacturonan,
Xyg = xyloglucan, AGP = arabinogalactan protein).

To complement the sugar composition and CoMPP analysis,

linkage analysis by permethylation was conducted. Linkage analysis of raw AIR generated unsatisfactory results with multiple interferring non-sugar peaks and hence Ba(OH)2 hydrolysed AIR was
used. Ba(OH)2 removes proteins by hydrolysis and as protein was
the single largest constituent of AIR this eliminated the interferring
peaks. It should be noted that this technique also solubilises polysaccharides and approximately 50% (w/w) of the arabinose and
mannose was removed by this method (results not shown) as mea-

sured by HPAEC-PAD. Furthermore peeling or solubilisation was

observed as indicated by the occurrence of t-glucose and 4-glucose
possibly originating from cellulose.
The linkage analysis (Table 3) showed that the majority of the
arabinose was linked via O-5, characteristic for pectic arabinan
and with attributes of a highly branched arabinan as indicated by
the presence of 3,5- and 2,3,5-linked arabinose. The linkages observed for mannose corresponds well with the mannose originating from high mannose type N-linked protein glycosylation
(Borisjuk et al., 2004; Lis and Sharon, 1978). The core GlcNAC
(N-acetylglucosamine) was not detected but Ba(OH)2 induced
deacylation cannot be excluded. The presence of the 4-galactose
and t-xylose linkages corresponds well with the presence of galactans and xylogalacturonan (Huisman et al., 2003). 2,4-Rhamnose
were detected in trace quantities, but due to the low quantity
the ratio between the two could not be accurately determined.
Side reactions of the Ba(OH)2 treatment prevents rm conclusions
on the cell wall data but it is evident that the majority of the
arabinose linkages resemble pectic arabinan linkages and the

Table 3
Linkage analysis of mature marama bean AIR after hydrolysis of proteins with
Ba(OH)2.
Sugar linkage

% area

Std. dev.

t-Arabinose
5-Arabinose
2,3,5-Arabinose
3,5-Arabinose
Sub-total

9.16
10.13
8.32
5.18
32.79

2.54
2.05
4.68
3.54
5.08

t-Mannose
2-Mannose
6-Mannose
3,6-Mannose
Sub-total

17.65
11.82
2.08
13.67
45.23

0.85
1.28
0.50
0.09
2.55

t-Galactose
4-Galactose
2,4-Galactose
Sub-total

3.02
3.34
1.81
8.16

0.86
0.03
0.22
1.10

t-Glucose
4-Glucose
Sub-total

2.42
9.88
12.30

0.84
0.34
1.18

t-Fucose
Sub-total

<1
<1

N/A
N/A

t-Xylose
Sub-total

1
1

N/A
N/A

2-Rhamnose
2,4-Rhamnose
Sub-total

<1
<1
1

N/A
N/A
N/A

Total

100.00

N/A, not applicable.

M.M. Mosele et al. / Phytochemistry 72 (2011) 14661472

60
TFA
KCl 1
1 M NaOH
6 M NaOH
R id 1
Residue

g//mg AIR
R

50

40

30

20

10

0
Fucose

Rhamnose Arabinose Galactose

Glucose

Xylose

Mannose

GalA

Sugars
Fig. 3. AIR sugars (lg/mg) of mature marama seeds with sequential extraction
(GalA = galacturonic acid).

majority of the mannose linkages resemble those present in protein glycosylation.

During maturation a decrease in content of non-cellulosic cell
wall components was observed. Of all the single sugars, only arabinose increased in abundance. The reason for the relative decrease
in non-cellulosic sugars from the AIR may be due to the increased
abundance of protein in the mature sample. The changes observed
in the composition of the cell wall sugars are probably due to cell

1469

wall adaptation to the maturation of the seed. In particular, the decrease in mannose content could be due to the crystallisation of
protein in protein bodies (Mosele et al., 2011) making it impregnable to the TFA used for sugar composition analysis similar to cellulose microbrils resistance to TFA hydrolysis.
To further characterise the cell wall, RG I was extracted from
mature samples using an enzyme extraction protocol according
to Harholt et al. (2006). Only low levels of RG I could be extracted
and the relative arabinose to rhamnose ratio of the extracted RG I
was lower than in the intact cell wall (results not shown). CoMPP
analysis of the CDTA extract gave a similar result (Fig. 2). The data
indicate the presence of two types of pectin: (1) a high HG pectin
readily extractable with CDTA and (2) a high arabinan pectin
requiring alkaline extraction procedures for solubilisation (Fig. 3).
In order to precisely quantify the extracted sugars and to complement the CoMMP analysis a modied sequential extraction was
carried out using the AIR fraction from mature seeds (Fig. 4). The
extraction protocol was designed to be different from the standardised CoMMP protocol in order to ensure extraction of specic
polymers in the different extraction steps. KCl in high concentrations can extract ionically bound cell wall proteins. As already
mentioned, both AGP and extensin were detected in the CoMPP
analysis and since both of them contain arabinose in their glycan
structures they could contribute to the high arabinose content of
the cell wall (Figs. 1 and 3). However, the quantity of arabinose extracted by KCl was low. The KCl fraction contained the majority of
the galactose found in the cell wall indicating its origin from AGP.

Mature marama bean

Alcohol insoluble residue

5 mg sample

TFA fraction

10 mg sample in dispersed in 0.2 M KCl, i.e. Complete tablet

with EDTA for 6 hours
hours.
Centrifuge.
Repeat overnight (14 hours).
Centrifuge.
Centrifuge.
Pool together first and second supernatant, freeze dry.

KCl fraction

Disperse pellets in 1 M NaOH containing 1 % NaBH4 for 6

hours.
Centrifuge.
Repeat overnight (14 hours)
hours).
Centrifuge.
Pool
first and
P l together
h fi
d second
d supernatant, ffreeze ddry.

1 M NaOH fraction

Disperse pellets in 6 M NaOH containing 1 % NaBH 4 for 6

hours.
hours
Centrifuge.
Repeat
R
overnight
i h ((14
14 hhours).
)
Centrifuge.
g
Pool together first and second supernatant, freeze dry.

Freeze dry remaining pellet

6MN
NaOH
OH ffraction
i

Residue fraction

Fig. 4. Sequential extraction of sugars from alcohol insoluble residue (AIR) of mature seeds.

1470

M.M. Mosele et al. / Phytochemistry 72 (2011) 14661472

Since only low levels of xylose were extracted, the glucose cannot
originate from xyloglucan. However, callose was observed in the
CoMPP analysis and could be the origin of the glucose (Fig. 2).
The next extraction was 1 M NaOH, and similar to both CDTA
and enzyme based extraction it proved unsuccessful in extracting
a high proportion of arabinose. The 1 M NaOH fraction contained
nearly 100% of the mannose but only low levels of other sugars.
Since legume seed proteins are extracted using alkaline extraction
procedures, the mannose may originate from extracted proteins
supporting the hypothesis that the mannose in the AIR fraction is
from protein mannosylation. A harsher extraction of 6 M NaOH
was also carried out which could release approximately 60% of
the arabinose. Some of the galacturonic acid also needed 6 M NaOH
for extraction, suggesting that this fraction of galacturonic acid is
linked together with the arabinans providing evidence that the arabinans observed are pectic arabinan. The low amounts of rhamnose
in the latter extractions indicate that the arabinans observed here
are long chained. Similar extraction protocols have been used on
other legume seeds in which the arabinose was also recalcitrant
to extraction (Shiga and Lajolo, 2006; Shiga et al., 2003, 2004),
but not to a degree as observed in marama beans. Using up to
4 M NaOH extractions Shiga and Lajolo (2006) could extract more
than 96% of the arabinose from the cell wall of common bean. Less
than 4% remained in the residue as unextractable arabinose, while
in marama bean around 40% arabinose remained in the residue.
Furthermore, there was a larger proportion of pectin in common
bean that was extractable with water or CDTA, than observed
when using enzyme mediated, 1 M NaOH or CDTA extraction of
marama bean. The data suggest the presence of more ramied or
branched arabinans in marama bean than in common bean.
Marama bean is normally not utilised and cooked like other
beans. They are instead roasted and consumed like peanuts. In
common bean a hard-to-cook phenomenon can be observed which
relates to the cell wall polymers extractability (Shiga et al., 2004).
Storage temperatures around or higher than room temperature
seem to induce a hardening of the bean that is correlated with
increasing recalcitrance of the cell wall polymers to extraction.
Since marama beans are stored at elevated temperatures in their
natural habitat the high recalcitrance could be an effect of acclimatisation to high temperatures.
Other legume seeds contain large amounts of arabinan but not
to the same level as that observed in marama bean. The increased
arabinan content of marama beans compared to common bean
could be the result of an adaption to the decrease in moisture level
of the seed. The moisture level of marama bean is low (ca. 5%)
(Holse et al., 2010; Mosele et al., 2011) compared to what is observed in for example, soya bean and other legumes (>8%) (Guo
et al., 2010; Khattab et al., 2009). It is hypothesised that arabinan
may play a role in plant tissues that undergo desiccation and the
plasticity of arabinan in the cell wall is tting with this hypothesis
(Gomez et al., 2009; Harholt et al., 2010).
The structure of marama bean arabinans is not fully elucidated
and the reason for their resistance to extraction remains elusive.
However, a tight association with cellulose could be anticipated
as 40% of arabinose was not extracted using 6 M NaOH. Normally,
6 M NaOH will readily solubilise almost all cell wall polysaccharides except cellulose. It is already known that galactan and arabinan can interact with cellulose microbrils (Zykwinska et al.,
2005). The interaction is strongest for debranched arabinan unlike
in the highly branched arabinan as observed in marama beans. The
specic degree of hydration in the marama bean cell wall is not
known. Since the water content is extremely low, it is likely that
regions of the cell wall are (semi) crystalline increasing its recalcitrance. In Chenopodiaceae covalent cross-linking is possible
through pectic side-chains terminated by feruloyl esters that can
dimerise (Fry, 1986; Rombouts and Thibault, 1986; Colquhoun

et al., 1994). This type or other types of cross-linking could also

be present in marama bean further increasing recalcitrance to
extraction. Further investigations are required in order to conrm
the presence of such cross-linking.
3. Conclusions
Investigation of the carbohydrate fraction of marama bean supports the presence of a seed composition similar to that of other
leguminous seeds, but with some atypical or unique characteristics. High levels of arabinose from arabinan and mannose from protein glycosylation were observed in the cell wall material. The
arabinose level observed was higher than that of other leguminous
species. The high arabinan content was not solubilised using standard pectin extraction methods, but was recalcitrant to extraction
as 60% of the arabinose was extractable. Even after 6 M NaOH
extraction a notable amount of arabinose could be detected in
the residue. Starch and soluble sugars were less than 1% (w/w).
4. Materials and methods
4.1. Sample preparation
Two different samples of marama bean were analysed constituting two different developmental stages; immature (unripe)
and the mature (fully ripe) stage of consumption. These samples
were harvested in 2008 from their natural habitat in the southern
region of Botswana. The samples were decorticated and the cotyledons were milled into our in a laboratory mill (IKA A10, Labortechnik, Staufen, Germany) for 15 s. Three samples of each our
were aliquoted and analysed in parallel.
4.2. Analysis of starch
Starch in the immature and mature bean our was quantied
by high-performance anion-exchange chromatography with
pulsed amperometric detection (HPAEC-PAD) as released glucose
after enzymatic hydrolysis of the starch. Defatted our (100 mg)
was extracted with 80% ethanol for 30 min at 80 C and the supernatants were discarded. Additional washing two times with 15 ml
0.1% sodium dodecyl sulphate (SDS) and thereafter with two times
15 ml water, with centrifugation between each step were done.
Thermostable a-amylase (Termamyl, Novozymes, DK) (300 U) in
50 mM (pH 7.0) MOPS buffer (4-morpholinepropanesulfonic acid
sodium salt, 5 mM, calcium chloride and 0.02% sodium azide)
was added to the samples, which were then incubated at 100 C
for 6 min. The temperature was then adjusted to 50 C, and sodium
acetate buffer (2.5 mM, pH 4.0) and 20 U of amyloglucosidase (Sigma, DK) were added. The samples were incubated for 30 min and
centrifuged and the supernatants were analysed for glucose content according to the method by Blennow et al. (1998). The starch
content was expressed as percentage of dry matter.
4.3. Analysis of soluble sugars
Soluble sugars in the immature and mature bean our were
quantied using HPAED-PAD. Defatted our (100 mg) was placed
in 15 ml conical tubes. The sugars were extracted with 80% ethanol
for 30 min at 80 C. The samples were centrifuged, and the supernatants recovered and dried overnight in a vacuum drier. The dry
residue of the supernatants were dissolved in 100 ll of water
and kept at 20 C prior to analysis. Sugar analysis was performed
as described by Lunde et al. (2008). The sugars were expressed in
nmol/g and ng/mg of dry matter.

M.M. Mosele et al. / Phytochemistry 72 (2011) 14661472

4.4. Isolation and identication of sugars from alcohol insoluble

residue (AIR)
A weight of 50 mg of immature and mature bean our was
placed in micro tubes. Extraction with benzene:ethanol (7:3 v/v,
1 ml) was carried out for 6 h. After centrifugation the pellets were
washed twice in acetone. They were further extracted with chloroform: methanol (3:2 v/v, 1 ml) for 72 h. After centrifugation, this
was followed by washing in 100% and 70% ethanol, and vacuum
drying. The residual material was designated as AIR. The samples
were added triuoro acetic acid (TFA) and hydrolysed for 1 h at
121 C, dried under vacuum and re-suspended in 1 ml water. They
were mixed, centrifuged and analysed by HPAEC-PAD according to
the method by bro et al. (2004). The sugars were expressed in lg/
mg of dry matter.
4.5. Comprehensive microarray polymer proling (CoMPP) analysis of
AIR
CoMPP was carried out as described by Moller et al. (2007) on
the mature bean our. Ten milligrams of AIR was sequentially extracted with 50 mM CDTA (1,2-diaminocyclohexane-N,N,N0 ,N0 -tetraacetic acid, pH 7.5, 4 M NaOH with 0.1% v/v NaBH4 and
cadoxen (31% (v/v) 1,2-diaminoethane with 0.78 M CdO). The extracts were printed in three dilutions and three replicates giving
a total of nine spots per sample. Primary antibodies were obtained
from PlantProbes, Leeds, UK or BioSupplies, Melbourne, Australia.
Secondary alkaline phosphatase-conjugated anti-rat or -mouse
antibodies were obtained from SigmaAldrich (Cat. Nos. A8438,
A3562 and A5588, respectively). Development of blots was performed using a BCIP/NBTC (5-bromo-4-chloro-30 -indolyphosphate/nitro-blue tetrazolium chloride)-based substrate. Arrays
were scanned, converted to 16 bit grey-scale TIFF format images
and spot signals quantied using the ImaGene 6.0 microarray analysis software (BioDiscovery, El Segundo, CA, USA). The extractions
were performed three times and the data presented as a heat map
is an average of these. The data was converted into the heat map
format using the online BAR heatmapper tool (http://bar.utoronto.ca/ntools/cgi-bin/ntools_heatmapper.cgi).
4.6. Ba(OH)2. hydrolysis of AIR
To prepare for linkage analysis (see Section 4.7) extensin (protein) was removed from mature bean our with Ba(OH)2. Without
the protein hydrolysis, imprecise results were obtained.
The samples were washed with water. They were then hydrolysed for 6 h at 120 C in 0.2 M Ba(OH)2 followed by neutralisation
with sulphuric acid to remove barium as precipitated BaCO3, as described by Fry (1988). The supernatants were removed and the
remaining pellets were analysed for linkages by permethylation.
4.7. Linkage analysis by permethylation
Freeze dried samples of Ba(OH)2 treated AIR were permethylated using the Hakomori procedure (Hakomori, 1964) as modied by
Huisman et al. (1998) and Verhoef et al. (2002). The permethylated
samples were hydrolysed in 2 M TFA containing 1 lmol myoinositol as internal standard, at 121 C for 1 h. The hydrolysed samples
were air-dried and subsequently reduced using sodium borohydride. The reduced, permethylated samples were converted into
alditol acetates and analysed using a Shimadzu gas chromatographelectron impact-mass spectrometer (GCEI-MS). The partly
methylated alditol acetates were separated on a 0.25  30 mm vitreous silica capillary column of SP-2330 (Supelco). The temperature was held at 80 C for 1 min upon injection, then
programmed from 80 to 170 C at 25 C per min, then to 210 C

1471

at 2 C per min, then to 240 C at 5 C per min, with a 10 min hold

at the upper temperature. Linkage composition was deduced from
both the relative retention times and the electron impact-mass
spectrum (EI-MS) (Carpita and Shea, 1989), and expressed as a percentage of dry matter.

4.8. Isolation and identication of sugars from AIR after sequential

extraction
Sequential extraction of AIR from mature bean our was done
in order to follow the extraction pattern of the different sugar fractions from AIR (Fig. 4). All fractions, except TFA fraction, were
washed several times in 70% ethanol, followed by centrifugation,
until the pH of each fraction was below 10. All fractions were dried
under vacuum, TFA hydrolysed and analysed for sugar composition
as described in Section 4.4, according to the method by bro et al.
(2004).

Acknowledgements
The authors wish to thank the National Food Technology Research Center (NFTRC) for providing marama beans. This work
was supported by the Government of Botswana, Ministry of Infrastructure, Science and Technology and the Danish Villum Kann
Rasmussen Foundation.

References
Aguilera, Y., Esteban, R.M., Benitez, V., Molla, E., Martin-Cabrejas, M.A., 2009a.
Starch, functional properties, and microstructural characteristics in chickpea
and lentil as affected by thermal processing. J. Agric. Food Chem. 57, 10682
10688.
Aguilera, Y., Martn-Cabrejas, M.A., Daz, M.F., Aguilera, Y., Bentez, V., Moll, E.,
Lo9 pez-Andru, F.J., Esteban, R.M., 2009b. Changes in carbohydrate fraction
during dehydration process of common legumes. J. Food Compos. Anal. 22, 678
683.
Amarteio, J.O., Moholo, D., 1998. The chemical composition of four legumes
consumed in Botswana. J. Food Compos. Anal. 11, 329332.
Avallone, R., Plessi, M., Baraldi, M., Monzani, A., 1997. Determination of chemical
composition of carob (Ceratonia siliqua): protein, fat, carbohydrates and tannins.
J. Food Compos. Anal. 10, 166172.
Blennow, A., Bay-Smidt, A.M., Olsen, C.E., Moller, B.L., 1998. Analysis of starchbound glucose 3-phosphate and glucose 6-phosphate using controlled acid
treatment combined with high-performance anion-exchange chromatography.
J. Chromatogr. A 829, 385391.
Borisjuk, L., Rolletschek, H., Radchuk, R., Weschke, W., Wobus, U., Weber, H., 2004.
Seed development and differentiation: a role for metabolic regulation. Plant
Biol. (Stuttg.) 6 (04), 375386.
Bower, N., Hertel, K., Oh, J., Storey, R., 1988. Nutritional evaluation of marama bean
(Tylosema esculentum, fabaceae): analysis of the seed. Econ. Bot. 42, 533540.
Caffall, K.H., Pattathil, S., Phillips, S.E., Hahn, M.G., Mohnen, D., 2009. Arabidopsis
thaliana T-DNA mutants implicate GAUT genes in the biosynthesis of pectin and
xylan in cell walls and seed testa. Mol. Plant 2, 10001014.
Carpita, N.C., Shea, E.M., 1989. Linkage structure of carbohydrates by gaschromatographymass spectrometry (GCMS) of partially methylated alditol
acetates. In: Biermann, C.J., McGinnis, G.D. (Eds.), Analysis of Carbohydrates by
GLC and MS. CRC Press, Inc., Boca Raton, pp. 157216.
Colquhoun, I.J., Ralet, M.C., Thibault, J.F., Faulds, C.B., Williamson, G., 1994. Structure
identication of feruloylated oligosaccharides from sugar-beet pulp by NMR
spectroscopy. Carbohydr. Res. 263, 243256.
Dalgetty, D.D., Baik, B.K., 2003. Isolation and characterization of cotyledon bers
from peas, lentils, and chickpeas. Cereal Chem. 80, 310315.
Fry, S.C., 1982. Phenolic components of the primary cell wall. Feruloylated
disaccharides of D-galactose and L-arabinose from spinach polysaccharide.
Biochem. J. 203, 493504.
Fry, S.C., 1986. Cross-linking of matrix polymers in the growing cell walls of
angiosperms. Annu. Rev. Plant Physiol. 37, 165186.
Fry, S.C., 1988. The Growing Plant Cell Wall: Chemical and Metabolic Analysis. The
Blackburn Press, Caldwell, NJ.
Gomez, L.D., Steele-King, C.G., Jones, L., Foster, J.M., Vuttipongchaikij, S., McQueenMason, S.J., 2009. Arabinan metabolism during seed development and
germination in arabidopsis. Mol. Plant 2, 966976.
Guo, W., Wang, S., Tiwari, G., Johnson, J.A., Tang, J., 2010. Temperature and moisture
dependent dielectric properties of legume our associated with dielectric
heating. LWT Food Sci. Technol. 43, 193201.

1472

M.M. Mosele et al. / Phytochemistry 72 (2011) 14661472

Hakomori, S., 1964. A rapid permethylation of glycolipid and polysaccharide

catalyzed by methylsulnyl carbanion in dimethylsulfoxide. J. Biochem. 55,
209212.
Harholt, J., Jensen, J.K., Srensen, S.O., Orla, C., Pauly, M., Scheller, H.V., 2006.
ARABINAN DEFICIENT 1 is a putative arabinosyltransferase involved in
biosynthesis of pectin arabinan in arabidopsis. Plant Physiol. 140, 4958.
Harholt, J., Suttangkakul, A., Scheller, H.V., 2010. Biosynthesis of pectin. Plant
Physiol. 153, 384395.
Holse, M., Husted, S., Hansen, ., 2010. Chemical composition of marama bean
(Tylosema esculentum) a wild African bean with unexploited potential. J. Food
Compos. Anal. 23 (6), 648657.
Holse, M., Larsen, F.H., Hansen, ., Engelsen, S.B., 2011. Characterization of marama
bean (Tylosema esculentum) by comparative spectroscopy: NMR, FT-Raman, FTIR and NIR. Food Res. Int. 44, 373384.
Hou, A., Chen, P., Shi, A., Zhang, B., Wang, J., 2009. Sugar variation in soybean seed
assessed with a rapid extraction and quantication method. Int. J. Agron. 2009,
article ID 484571.
Huisman, M.M.H., Schols, H.A., Voragen, A.G.J., 1998. Cell wall polysaccharides from
soybean (Glycine max) meal. Isolation and characterisation. Carbohydr. Polym.
37, 8795.
Huisman, M.M.H., Schols, H.A., Voragen, A.G.J., 2003. Pectic substances from
soybean cell walls distinguish themselves from other plant cell wall pectins.
In: Voragen, F., Schols, H., Visser, R. (Eds.), Advances in Pectin and Pectinase
Research. Kluwer Academic Publishers, Netherlands, pp. 159168.
Jensen, J.K., Sorensen, S.O., Harholt, J., Geshi, N., Sakuragi, Y., Moller, I., Zandleven, J.,
Bernal, A.J., Jensen, N.B., Sorensen, C., Pauly, M., Beldman, G., Willats, W.G.T.,
Scheller, H.V., 2008. Identication of a xylogalacturonan xylosyltransferase
involved in pectin biosynthesis in arabidopsis. Plant Cell 20, 12891302.
Karr-Lilienthal, L.K., Kadzere, C.T., Grieshop, C.M., Fahey Jr., G.C., 2005. Chemical and
nutritional properties of soybean carbohydrates as related to non-ruminants: a
review. Livest. Prod. Sci. 97, 112.
Ketshajwang, K.K., Holmback, J., Yeboah, S.O., 1998. Quality and compositional
studies of some edible Leguminosae seed oils in Botswana. J. Am. Oil Chem. Soc.
75, 741743.
Khattab, R.Y., Arnteld, S.D., Nyachoti, C.M., 2009. Nutritional quality of legume
seeds as affected by some physical treatments: Part 1. Protein quality
evaluation. LWT Food Sci. Technol. 42, 11071112.
Khurana, H.K., Kanawjia, S.K., 2007. Recent trends in development of fermented
milks. Curr. Nutr. Food Sci. 3, 91108.
Lis, H., Sharon, N., 1978. Soybean agglutinin a plant glycoprotein. Structure of the
carbohydrate unit. J. Biol. Chem. 253, 34683476.
Lunde, C., Zygadlo, A., Simonsen, H.T., Nielsen, P.L., Blennow, A., Haldrup, A., 2008.
Sulfur starvation in rice: the effect on photosynthesis, carbohydrate
metabolism, and oxidative stress protective pathways. Physiol. Plant. 134,
508521.
Martn-Cabrejas, M.A., Daz, M.F., Aguilera, Y., Bentez, V., Moll, E., Esteban, R.M.,
2008. Inuence of germination on the soluble carbohydrates and dietary bre
fractions in non-conventional legumes. Food Chem. 107, 10451052.
Mitei, Y.C., Ngila, J.C., Yeboah, S.O., Wessjohann, L., Schmidt, J., 2008. NMR, GCMS
and ESI-FTICR-MS proling of fatty acids and triacylglycerols in some Botswana
seed oils. J. Am. Oil Chem. Soc. 85, 10211032.

Mitei, Y.C., Ngila, J.C., Yeboah, S.O., Wessjohann, L., Schmidt, J., 2009. Proling of
phytosterols, tocopherols and tocotrienols in selected seed oils from Botswana
by GCMS and HPLC. J. Am. Oil Chem. Soc. 86, 617625.
Moller, I., Sorensen, I., Bernal, A.J., Blaukopf, C., Lee, K., Obro, J., Pettolino, F., Roberts,
A., Mikkelsen, J.D., Knox, J.P., Bacic, A., Willats, W.G.T., 2007. High-throughput
mapping of cell-wall polymers within and between plants using novel
microarrays. Plant J. 50, 11181128.
Mosele, M.M., Hansen, .S., Hansen, M., Schulz, A., Martens, H.J., 2011. Proximate
composition, histochemical analysis and ultrastructural localisation of nutrients
in immature and mature seeds of marama bean (Tylosema esculentum) an
underutilised legume. Food Chem. 127, 15551561.
Mullin, W.J., Xu, W., 2000. A study of the intervarietal differences of cotyledon and
seed coat carbohydrates in soybeans. Food Res. Int. 33, 883891.
Nakamura, A., Furuta, H., Maeda, H., Takao, T., Nagamatsu, Y., 2002. Analysis of the
molecular construction of xylogalacturonan isolated from soluble soybean
polysaccharides. Biosci. Biotechnol. Biochem. 66, 11551158.
National Research Council, 1979. Marama Bean, in Tropical Dry Pulses: Resources
for the Future. National Academy of Sciences, Washington, DC, pp. 6874.
bro, J., Harholt, J., Scheller, H.V., Orla, C., 2004. Rhamnogalacturonan I in Solanum
tuberosum tubers contains complex arabinogalactan structures. Phytochemistry
65, 14291438.
Reddy, N.R., Pierson, M.D., Sathe, S.K., Salunkhe, D.K., 1984. Chemical, nutritional
and physiological aspects of dry bean carbohydrates a review. Food Chem. 13,
2568.
Rombouts, F.M., Thibault, J.F., 1986. Feruloylated pectic substances from sugar-beet
pulp. Carbohydr. Res. 154, 177187.
Saldivar, X., Wang, Y., Chen, P., Hou, A., 2011. Changes in chemical composition
during soybean seed development. Food Chem. 124, 13691375.
Scheller, H.V., Jensen, J.K., Srensen, S.O., Harholt, J., Geshi, N., 2007. Biosynthesis of
pectin. Physiol. Plant. 129, 283295.
Shiga, T.M., Lajolo, F.M., 2006. Cell wall polysaccharides of common beans
(Phaseolus vulgaris L.) composition and structure. Carbohydr. Polym. 63, 112.
Shiga, T.M., Lajolo, F.M., Filisetti, T.M.C.C., 2003. Cell wall polysaccharides of
common beans (Phaseolus vulgaris L.). Cinc. Tecnol. Aliment. 23, 141148.
Shiga, T.M., Lajolo, F.M., Filisetti, T.M.C.C., 2004. Changes in the cell wall
polysaccharides during storage and hardening of beans. Food Chem. 84, 5364.
Verhoef, R., de Waard, P., Schols, H.A., Rtt, M., Siika-aho, M., Voragen, A.G.J., 2002.
Structural elucidation of the EPS of slime producing Brevundimonas vesicularis
sp. isolated from a paper machine. Carbohydr. Res. 337, 18211831.
Viarta, S.C., Molina, O.E., Figueroa, L.I.C., Faria, J.I., 2006. A further insight into the
applications of exopolysaccharides from Sclerotium rolfsii. Food Hydrocolloids
20, 619629.
Vincken, J.P., Schols, H.A., Oomen, R.J.F.J., McCann, M.C., Ulvskov, P., Voragen, A.G.J.,
Visser, R.G.F., 2003. If homogalacturonan were a side chain of
rhamnogalacturonan: I. Implications for cell wall architecture. Plant Physiol.
132, 17811789.
Zykwinska, A.W., Ralet, M.C.J., Garnier, C.D., Thibault, J.F.J., 2005. Evidence for
in vitro binding of pectin side chains to cellulose. Plant Physiol. 139, 397407.